1. Field of the Invention
The present invention relates to communication systems. More specifically, the present invention relates to a system and method for frequency offsetting of information communicated in multiple input/multiple output-based communication systems.
2. Discussion of Related Art
In wireless communication systems, efficient data transmission may be achieved using a multiple input/multiple output system (“MIMO” or “MIMO system”). At its simplest, a MIMO system employs a single transmitter or a plurality of chained transmitters (“chain” or “chains”) associated with multiple physical transmitting antennas to send simultaneously multiple data streams (“signals”) through a radio channel. The multiple data streams are received by multiple receiving antennas associated with a single receiver or a plurality of chained receivers (“chain” or “chains”).
This system results in better spatial utilization of the radio channel bandwidth. In turn, higher throughput, improved link reliability, and improved spectral efficiency are achieved. A MIMO channel includes channel impulse responses or channel coefficients in the flat fading case between different pairs of transmitting and receiving antennas. As is known in the art, a MIMO system may be modeled asy=Hx+n  (Equation 1)where x and y are the transmit and receive sign vectors, respectively, n is the channel noise vector, and H is channel matrix.
MIMO systems are most useful in indoor environments where walls, ceilings, and furniture provide a rich multi-path environment, such that the channel matrix allows for multiple independent and orthogonal impulse responses or spatial signatures. In such an environment, the MIMO technology is able to transmit multiple parallel and independent data streams relying on the orthogonal elements of the channel matrix. MIMO systems deployed in highly scattering environments produce high ranked H matrices resulting in higher MIMO capacities even when low correlated antennas are used.
MIMO systems which have been developed for 4G IEEE 802.16e WiMAX systems, have been optimized with two central goals in mind: (1) to maximize/optimize spectral efficiency; and (2) to dynamically achieve improvements in coverage gain or reach by reducing spectral efficiency.
For cellular vendors, spectrum is a precious and limited resource where revenue is defined largely as a function of system capacity and throughput. Spectral efficiency is therefore of paramount importance for these networks where revenue is measured as a functions of carried bandwidth. A significant portion of a cellular provider's operating expenses are from the monthly leasing fees for each cell site.
Maintaining existing cell site coverage is also critical since ubiquity of service is a requirement for any 4G wireless network, and yet the increased delivered channel bandwidth would have reduced link budgets and therefore smaller cell sizes. Cellular providers rely on MIMO technology and the ability to tradeoff capacity for reach at the cell edge to maintain the current cell coverage.
Cellular providers, which are by far the largest economic force driving the advancement of MIMO systems, have maintained the industry focus on spectral efficiency and dynamic reach tradeoff, as well as innovative antenna systems at the base station (BS) and station set (SS) equipment. Those in the WiMAX industry are familiar with “Matrix A” for coverage gain—where a single data stream is transmitted in parallel over two independent transmitter-antenna-receiver paths using space time block codes (STBC) to encode the two streams such that they are orthogonal to each other, thereby improving the signal-to-noise ratio (SNR) at the receiver, resulting in increased cell radius. “Matrix B” was developed for capacity increases which use the spatial multiplexing of MIMO to transmit independent data streams with throughput capacity limited only by the rank of the H matrix and the local noise floor characteristics.
MIMO systems which have been developed for IEEE 802.11n wideband local area network (WLAN) systems have been optimized with the same two central goals driven by the cellular industry's 4G systems—maximizing spectral efficiency and capacity; and optimizing coverage. However, WLAN vendors have overriding industry requirements of solution size, power and cost, as these chipsets are now being embedded in every laptop PC sold as well as in all of the new cellular telephones and personal digital assistants (PDAs). WLAN solution providers have made incredible gains since the first IEEE 802.11b radios were introduced less than a decade ago. WLAN solutions have progressed in the areas of capacity and range as the WiFi standard has evolved from the 11 Mbps systems based on IEEE 802.11b with an effective throughput of 6 Mbps to the 54 Mbps OFDM systems of IEEE 802.11a and IEEE 802.11g with an effective throughput in the range of 25 Mbps. The introduction of IEEE 802.11n with MIMO has been demonstrated to show peak throughputs as high as 300 Mbps and effective throughputs equivalent to 100 Mbps for most house hold applications where MIMO technology is able to perform well.
The same WLAN solutions providers have focused their efforts on cost reduction, by fully integrating chips and radio frequency transmitters to the point that a single chip is able to support all software functions as well as transmit and receive with a zero-IF (ZIF) architecture. Power reduction of these single chip solutions has allowed for a limited mini-PCI power budget of approximately 3W, supporting a 3×3 IEEE MIMO 802.11n protocol with relatively high powered transmitters in the range of 17 dBm per channel, as high as 20 dBm with the typical <3 dBi gain antennas used in laptops or for WLAN consumer access points. The same WLAN vendors have been less interested in spectral efficiency and have allowed channel sizes to increase from 20 MHz bandwidths to 40 MHz bandwidths.
While MIMO systems operate best in indoor or highly scattering environments which produces high ranked H matrices resulting in higher MIMO capacities, cellular systems employing MIMO are deployed outdoors and often in line-of-sight (LoS) or near LoS (NLoS) applications. High gain antennas—between 10 dBi to 30 dBi—may be used for long distance point-to-point links. It is not as well known in the industry that radio scattering, also called multipath interference, is related directly to the beamwidth of the antenna such that high gain narrow beam antennas will see less multipath interference as do lower gain wide beamwidth antennas. This less obvious fact makes logical sense, as high gain antennas have a narrow antenna beamwidth and therefore a small aperture capable of receiving strong radio signals. Signals received by such a narrow aperture will, in fact, have traveled similar distances resulting in minimal multipath interference. Another way to understand this fact is by considering the reception of a high energy RF “impulse” generated by a transmitter and received by a receiving antenna. The impulse will bounce off of many obstacles, arriving with at the receiving antenna as an impulse response. A receiving antenna with a narrow aperture pointed directly at the source of the impulse will reject any of the longer delay echoes of the impulse, which tend to come from sources which are not directly in-line with the transmitter. Thus, outdoor high gain directional point-to-point MIMO systems cannot rely on multipath dispersion or radio scattering as a means of increasing the rank of the spatial H matrix; however, other means including spatial separation and polarization diversity are possible.
Cellular vendors have long relied on spatial separation to achieve independence of the multipath reflections for antenna diversity receivers in outdoor environments. Many papers have been written regarding spatial separation of receiving antennas. In general, when the receiving antenna is mounted at a low height and is close to reflecting and scattering objects, then a very small separation in the range of one half of a wavelength or just a few inches is required to achieve channel multipath independence. However, when the receiving antennas are mounted high on towers or rooftops, as is most often the case, then small separations have no significant reduction of the correlation of the multipath signatures, and larger separations, on the order of meters, must be used to gain independence of the radio channels. Most antenna systems mounted on rooftops, cell towers, and other elevated structures separate diversity receive antennas by 2 meters or more to achieve path independence to realize gains from antenna diversity. MIMO access systems can rely on the same antenna separation to improve overall throughput.
Polarization diversity can also be used to achieve independence of the radio channel. Most IEEE 802.16e MIMO systems currently being deployed use slant diversity in each of the antennas, and three or more antennas separated by 2 meters each can achieve gains of beam steering as well as a high order channel matrix H. Unfortunately, wireless backhaul networks cannot afford to have multiple receiver antennas separated by two or more meters because of the existing lease agreements for antenna attachment. These lease agreements typically limit an antenna to be less than 1 ft×1 ft×4 ft in total size, including the transceiver equipment itself.
Moreover, cellular providers have further restricted equipment manufacturers of point-to-point radio equipment for backhaul purposes to be less than 1 ft×1 ft×1 ft in total size, and this has become an industry “norm” for such equipment. This restriction effectively limits the allowed antenna gain, but allows for antenna diversity to be used to achieve independence of the MIMO paths and allows for as high as a 2×2 matrix.
Antenna polarization diversity works well for links which are LoS with no possibility of obstacles within the Fresnel zone of the radio. In such cases, the MIMO gains can be determined a priori so that the network planner is able to accurately define how many radio links and their specified bandwidth that will be achieved using the 1 ft×1 ft×1 ft transceivers.
For the case of non-LoS or near LoS point-to-point links that experience time varying reflections, MIMO gains are less well characterized and may only be a fraction of the maximum possible throughput. As an example, a 2×2 MIMO transmission formed using antenna polarization diversity will see continuous polarization rotations if the signals pass through wet foliage, such as trees, after a rainfall. The presence of a few trees in the Fresnel zone typically results in a 10 dB reduction in transmitter signal strength, a condition such that even a light breeze can change the propagation channel more quickly than the hardware algorithms are able to handle and update the channel matrix “H” to maintain full throughput. As a result, for these types of links, the MIMO gains are difficult to quantify for network capacity planning.
Given the difficulties in quantifying the capacity for a 2×2 MIMO point-to-point radio link, the effort becomes even more challenging with a 3×3 or 4×4 MIMO solution. These higher MIMO solutions under ideal conditions deliver significantly higher capacity than a non-MIMO solution, yet their effectiveness is governed by site-specific issues of LoS and near-LoS path characteristics. There are no documented procedures or guidelines which specify assured/minimum MIMO gains for a given antenna separation; therefore, the installer and network planner has no accurate means to know before deploying the MIMO radios what the links capacity will be.
Finally, even under the best conditions of LoS and antenna isolation and separation, interference in an unlicensed band is always an issue. In many environments, unlicensed band interference can be described as a general noise floor, driven by tens, hundreds, or thousands of individual and geographically dispersed sources, where typically just a few sources dominate.
The vast majority of interference sources tend to be in a fixed location—e.g., radiation from microwave ovens or DECT wireless phones, or even pinball machines. Some are mobile, such as Bluetooth devices or laptops. In general, for outdoor point-to-point networks, the noise floor tends to be static in nature, but with sudden changes when a mobile source is introduced near to the point-to-point microwave link. These sources are not well handled by MIMO radio links which are channel specific and are thus affected on all MIMO paths by a single interference source.
Thus, there is a need for an improved MIMO system that provides for greater bandwidth and greater assured reliability. There is also a need for a MIMO system that requires limited antennas to permit usability in limited physical spaces.
In MIMO based technologies such as IEEE 802.11n Wi-Fi or IEEE 802.16e WiMAX, the transmitters have been designed to generate multiple output data streams using common crystal oscillators for the baseband and common local oscillator(s) (LOs) for the conversion to radio frequency (RF), and where the final RF signals are at the same frequencies. The phase variations present in the baseband and LO circuits will be seen equally on all of the MIMO RF signals so that a MIMO receiver can recover timing from any one of the MIMO RF signals and apply that timing to all of the other streams.
For example, a MIMO transmitter generates multiple MIMO RF signals at 5 GHz using a crystal with a +10 parts per million (ppm) error. The RF signals are transmitted over the air at 5 GHz+10 ppm=5,000,050,000 Hz. The MIMO receiver would receive the multiple RF signals using a crystal with a −10 ppm error, so that the down conversion would be with a 4,999,950,000 Hz signal. The resulting signal at baseband would have a frequency error of 100,000 Hz=100 kHz on all MIMO streams, which is easily tracked and removed by the timing recovery function on any one of the recovered MIMO signals.
However, if the system shifts frequency of the MIMO streams to different radio frequencies, a new problem occurs. When these different RF streams are down converted, the resulting errors (as measured in Hz) will be different for each MIMO stream generated from the various RF signals. For example, using the frequency shifter, a 2×2 MIMO transmitter may generate two MIMO RF signals at 5 GHz and 6 GHz using a crystal with a +10 ppm error. These RF signals will be transmitted over the air at 5 GHz+10 ppm=5,000,050,000 Hz and 6 GHz+10 ppm=6,000,060,000 Hz. The MIMO receiver will receive the two RF signals using a crystal with a −10 ppm error, so that the down conversion will be with a 4,999,950,000 Hz signal and a 5,999,940,000 Hz signal on the first and second MIMO streams respectively. The resulting signals at baseband will have a frequency error of 100 kHz for the first MIMO stream and 120 kHz for the second MIMO stream. If the receiver derives its timing recovery from the first MIMO stream, then the second MIMO stream will have an error of 120 kHz−100 kHz=20 kHz. This 20 kHz error, as seen on a 250 us packet, will appear as 5 complete rotations of the OFDM constellation, thus making timing recovery impossible for any of the modulation rates. It is noted that the +10 ppm and −10 ppm frequency errors used above are shown only for purposes of a simplified example calculation. Typical WLAN devices use crystals having frequency errors within the range of +/−20 ppm. Further, the above example assumes that the MIMO receiver derives its timing recovery from a single stream. If the MIMO receiver derives its timing on a per-stream basis, then the exemplary 20 kHz frequency error may be inconsequential if the MIMO receiver can support relatively large frequency variations.
Accordingly, there is a need to address the problem of different down conversion frequency errors for frequency-shifted RF streams in MIMO systems.